Life and Graph Paper

[A warning:  this is going to feel like being on a long bus ride with someone who can't stop talking about some weird, inaccessible thing which that person finds absolutely fascinating.  I've been rewriting this stupid entry for six weeks and this is as conversational as it's going to get.  I don't think I'm doing justice to the subject matter or to those of you who are nice enough to check in here and look around, but it was either this or the 400 insomnious words I wrote early one morning about Steve Bannon's hair.]

It's weird, the things that stay with you.

In October of 1984, I read an article in Omni  magazine about a game developed by an English mathematician. That article contained a single sentence, the philosophical ramifications of which I'm still feeling thirty years later, which is a lot of bang for the buck for any magazine, much less one I'd bought just because it had a lot of pictures of plasma generators in it.

In 1970, John Conway began experimenting with cellular automation -- a concept that had existed since the late forties but enjoyed little popular interest for reasons that are about to become bluntly apparent.

Cellular automation (CA) is simply applying a set of rules to an array of cells so that the condition of any one cell is influenced by the cells around it. So, you'd have a sheet of graph paper, you'd designate some squares as being in some specific state, you'd get a second sheet of graph paper and apply your rules to calculate the influence of the cells on one another. The second sheet represents the second "generation" of cells. And you'd repeat this for as long as the rules continued to change the state of the array.  Sound fun?

We're already in deep water here, so let's go ahead and touch the bottom of the pool before we come up for air.

As an illustration of CA's limited early appeal, consider that John von Neumann, one of its inventors, seems to have spent the early fifties using it as the world's least sexy way to model reproduction in biological systems. He designed three dimensional cellular arrays (bleh) in which cells could exist in twenty nine different states (which makes my head hurt) as governed by differential equations (I don't know what those are). I'm sure he was having a blast with all of that but, wow. No thank you.

In 1969, John Conway's interest in CA led to his developing a set of rules that produced interesting results and, at the same time, were easy to understand. The net result being that a lot of people began playing with and exploring cellular arrays that were governed by his rules, which were applied to a nice, manageable, hobbyist-friendly two dimensional array. Those rules are these:

• Cells exist in only two states, either alive or dead.
• A dead cell becomes alive if three of its neighboring cells are alive.
• A living cell dies if less than two or more than three of its neighboring cells are alive.

Here's an illustration from that article which traces the development of a shape called the "T tetromino," the name of which communicates with equal effectiveness the shape of the configuration as well as how annoying it must have been to share a dorm room, work space or cafeteria table with anyone who was so deeply into this.

There's a recursion, by the way, in the last two steps.  Here's a written description of what's going on.

The four sets of three connected red dots designate these "cells" as "alive." Apply the rules, you'll see that the two squares on the ends, being connected to only one "living cell," will "die" in the next "generation", the middle "cell," being connected to two "living" cells, survives to the next "generation", and that there are two currently "dead" cells that will "born" because they are adjacent to three "living" cells. This results in three "alive" cells being arranged in a vertical/horizontal line which, obviously, will revert to the original configuration in the next generation, making this shape a "blinker" and some people find all of this very "interesting."

I'm not sure I actually would have been one of those interested people if the article which described Conway's game ("Life," as he came to call it) hadn't contained the following quote from a book Conway co-authored entitled Winning Ways (Academic Press, 1982). The first time I read it, I felt like someone had pulled my brain out through my eye. 

"It's probable, given a large enough Life space, initially in a random state, that after a long time, intelligent, self-replicating animals will emerge and populate some part of the space."

The implications of that statement hit me very hard because it was the first time I'd ever considered what it was that defines something as being alive.

There's that philosophical safe haven of life being inherently biological, but relying on such a narrow definition seems arbitrary when we're just thinking about a series of reactions.  The temptation to speculate that it's possible for "living creatures" to spawn and begin conducting their own business in this realm was hard to dismiss amid the increasing availability of computers allowing Life enthusiasts to generate large, random colonies of shapes and watch the consequent battle between order and mayhem.  Some of the simulation's emergent shapes do actually crawl around on the grid in a cohesive way.  Some self-replicate.  Some produce parts that collide with other parts generated by different clusters and they, in turn merge into configurations that crawl off to do Tron-knows-what in some other part of the grid.  It's weird.

You see this sort of thing happening in Minecraft now, but it was probably significantly trippy the first time someone designed configurations of cells that can be employed to create logic gates which will work identically to (although, for now, more slowly than) the two billion transistors that do the heavy lifting inside every iPhone 6.  Putting aside that the whole system is deterministic (that's an existential panic attack for another day) it really kind of does seem like there is some way that these little blinky so-and-so's might be considered to be, on some level, awake.

But, this is the thing that really creeped me out at fifteen.

If any of those configurations prove to meet the criteria for being alive, it's not because they exist now in digital environments and are propelled by CPU cycles. Bear in mind that Conway and early devotees of this game were executing it by hand. Every behavior exists in some theoretical realm waiting to be described by either a blazingly fast supercomputer or a compulsively diligent math hobbyist with cases of pre-sharpened pencils and stacks upon stacks of graph paper.  If it's possible for configurations of dots on paper can be considered alive, "alive in a way that a virus is not," according to someone named Poundstone who is quoted in the article and is undoubtedly smarter than I am by a bunch, then I am never... ever... going to Office Max again.

My reacquaintance with the Omni article happened around the same time I heard something really interesting on Tested's Still Untitled podcast:  Photosynthesis and combustion are the same process running in two different directions. One converts hydrocarbons and oxygen into carbon dioxide, light and heat via an exothermic reaction, the other uses an endothermic reaction to convert carbon dioxide, light and heat into hydrocarbons and oxygen. In summary, to quote Adam Savage, the speaker who had just laid this at the listener's feet, "So the opposite of fire... is plants."

That's interesting and all, but plants are alive and fire, by most conventional definitions, is not.

Consider, though, that one of most vague definitions of life, "matter capable of extracting energy from the environment for replication," seems so broadly inclusive as to bestow the quality of life to fire. It also, kind of, in a way, if you step back a couple of feet and squint, describes the existence of these infuriating little theoretical squares within cellular automata if you categorize the constant intellectual curiosity of a hundred thousand people as "energy," which is not a case I'd want to argue against. 

The capacity exists now for arrays large enough to accommodate Conway's imagined creatures, and so far, nothing has tapped on the glass.  Online resources for the game are vast, dizzying videos of intricate configurations are available on Youtube, as are interviews with John Conway -- who is glad that people found this pursuit interesting, but laments that is has overshadowed his other accomplishments.  Maybe that's the infestation he was really sniffing out thirty years ago.





Appendix:  

[Here's a timeline of my personal history with Life.]

October 1984 -- I read the aforementioned article in Omni 

Early 1985 -- I attempt to re-read article and realize I've misplaced the magazine

Rest of 1985 -- I search the house for the magazine.

1986-1990 -- I browse libraries and used book sales for said issue without success, although I do accumulate other back issues of Omni. So, yay.

1990 -- Someone gives me a copy of Applesoft BASIC and I take a shot at coding a program that will actually run these rules. I eventually get something to crank out large sets of coordinates and cell-states, but every attempt to write a program that turns all of this into observable graphics fails on every level.  Well, mostly just the one level.  The one where they don't run.  We'll never know if they could have failed on some other more interesting level.  I like to think they would have, but I'm kind of a dreamer.

1993 -- I get a job selling computers and discover that there is a version of Conway's Life included on the third Microsoft Entertainment Pack. reading the Help file, I discover the rules I had implemented in my version were wrong, invalidating all of my incomplete programming and also all of the cycles [which I had played out on graph paper and was, yes, a little cumbersome but totally doable -- not that you'd know or care to find out].

1994 -- I find some discussions and software on CompuServe (the online service I joined because I'd read about it in, as it turns out, that same issue of Omni). 

1995 -- Omni stops publication.

2013 -- The entire run of Omni becomes available in PDF format at archive.org.

2015 -- The entire run of Omni becomes unavailable in PDF format at archive.org.

Like, last Thursday or something -- I kind of, you know... Pirate Bay that archive. And here we are.